accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Eric Newton (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-2866) Default WAL size should be based on HDFS Block Size
Date Fri, 06 Jun 2014 17:38:01 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-2866?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14020107#comment-14020107
] 

Eric Newton commented on ACCUMULO-2866:
---------------------------------------

In the normal case, WALs are written out and never re-read.  But, if a tablet is lost, the
contents of memory have to be recovered.  The master coordinates getting the WAL sorted into
a recovery file.  The tservers use the sorted recovery file to re-load the data that was in
memory, and not yet flushed to disk.

WAL file = not sorted, rarely read
recovery file = sorted, and read by all recovering tablets/servers.

> Default WAL size should be based on HDFS Block Size
> ---------------------------------------------------
>
>                 Key: ACCUMULO-2866
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-2866
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: tserver
>            Reporter: Mike Drob
>            Priority: Minor
>              Labels: wal
>             Fix For: 1.7.0
>
>
> HBase automatically defaults their WAL size to 0.95 * HDFS block size. This makes a lot
of sense from a resource management perspective, and we should do the same.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message